Nvidia has announced a new round of investment in Thinking Machines Lab, an artificial intelligence (AI) company founded by former OpenAI executive Mira Murati, and will provide the startup with chips to train and run its AI models.

Both companies stated in a press release on Tuesday that, under a multi-year agreement, Thinking Machines Lab will use Nvidia’s upcoming Vera Rubin AI acceleration chips. These chips are expected to be deployed early next year and will provide Thinking Machines Lab with at least 1 gigawatt of computing power (equivalent to the electricity consumption of approximately 750,000 households).

Nvidia had previously invested in Thinking Machines Lab last year. The specific terms of this investment were not disclosed, and it was not clarified whether the funding would be in cash, chips, or a combination of both. Nvidia referred to this as a “significant investment.” A spokesperson for Thinking Machines Lab declined to provide further details, and Nvidia has not yet responded to requests for comment.

As the world’s most valuable company, Nvidia has been involved in a series of investment deals in recent months. The chip giant is leveraging its resources to drive AI adoption across various industries and help fuel what it describes as a “new industrial revolution.” However, these investments have drawn attention due to the closed-loop nature of the model—Nvidia is investing in its own customers.

According to reports from November of last year, Thinking Machines Lab had been seeking a new round of funding, targeting a valuation of $50 billion. If achieved, this would represent a fourfold increase from its previous valuation of $12 billion in July, when the company completed a $2 billion financing round at that valuation.

Murati stated in a press release, “This partnership will accelerate our goal of building AI that users can truly shape and own, which will, in turn, unlock human potential.”

Murati, who previously served as OpenAI’s Chief Technology Officer, has led Thinking Machines Lab to recruit dozens of employees from OpenAI. The company launched its first product, Tinker, in October of last year. Tinker is designed to help users optimize large language models—this is the underlying technology behind chatbots like ChatGPT.

However, in recent months, Thinking Machines Lab has faced talent return, with several employees, including its CTO, rejoining OpenAI.

On March 2, U.S. stocks in the optical communication sector rose against the market trend. NVIDIA announced strategic multi-year partnerships with two leading companies in the optical communications field, Lumentum and Coherent, investing $2 billion in each, for a total of $4 billion. This investment will focus on the research and development of advanced optical technologies and manufacturing, accelerating the large-scale development of the next-generation AI infrastructure. It further strengthens NVIDIA’s global leadership in AI and accelerated computing.

The partnerships are non-exclusive agreements, each including billions of dollars in product procurement commitments, as well as capacity usage and priority rights for future advanced lasers, optical networks, products, and components.

NVIDIA’s investment will primarily support the research and development work of Lumentum and Coherent, future capacity expansion, and daily operations, while also helping the companies increase their U.S. domestic manufacturing capabilities. Lumentum will build a new wafer fab, and Coherent will expand its domestic manufacturing presence.

Optical interconnect technology and advanced packaging integration are the key foundations for continuously expanding AI computing factories and achieving ultra-high bandwidth and energy-efficient interconnectivity. These are critical components of the next-generation AI infrastructure. Through this cooperation, NVIDIA aims to leverage its technological and market advantages in AI, accelerated computing, and networking, while combining Lumentum’s strengths in optical and photonics technology and Coherent’s expertise in optical innovation and advanced manufacturing. The goal is to push breakthroughs in cutting-edge fields such as silicon photonics technology and support both partners in increasing their capacity and research investments to meet the construction needs of global next-generation AI data centers.

Jensen Huang, founder and CEO of NVIDIA, stated, “Artificial intelligence is reshaping computing models and driving the largest-scale infrastructure buildout in history. This partnership with two leading companies will help NVIDIA develop more advanced silicon photonics technology and accelerate AI infrastructure breakthroughs in scale, speed, and energy efficiency to create gigawatt-level next-generation AI computing factories.”

Jim Anderson, CEO of Coherent, commented, “This strategic partnership reaffirms Coherent’s critical role as a core enabler of next-generation artificial intelligence data center infrastructure. We are honored to deepen this 20-year-long partnership with NVIDIA, providing support for a wide range of products to help them build future AI data centers.”

Michael Helston, CEO of Lumentum, said, “This multi-year strategic agreement demonstrates our mutual commitment to advancing optical technology innovations that will become the driving force of next-generation AI infrastructure. To support this collaboration, we will invest in building a new manufacturing plant to increase capacity and accelerate technological innovation. We look forward to working with NVIDIA to continuously break through technical boundaries and unlock more possibilities for future AI optical architectures.”

Lumentum is a global leader in optical and photonics technology, providing core support for AI and cloud computing network infrastructure. It is headquartered in San Jose, California. Coherent, founded in 1971, is a leader in photonics, with operations in over 20 countries, providing world-leading photonics technology for data centers and communications.